Merge pull request #21 from oreillymedia/SPIDR-1451

SPIDR-1451 keep highest weight, de-dupe on text AND context

pom.xml

@@ -2,7 +2,7 @@
   <modelVersion>4.0.0</modelVersion>
   <groupId>ifpress</groupId>
   <artifactId>ifpress-solr-plugin</artifactId>
-  <version>1.5.7</version>
+  <version>1.5.8</version>
   <name>ifpress solr plugin</name>
   <description>Contains plugins to be installed in the solr server</description>
   <dependencies>

src/main/java/com/ifactory/press/db/solr/spelling/suggest/SafariInfixSuggester.java

@@ -12,14 +12,13 @@
 import org.apache.lucene.search.suggest.analyzing.AnalyzingInfixSuggester;
 import org.apache.lucene.store.Directory;
 import org.apache.lucene.util.BytesRef;
-import org.apache.lucene.util.Version;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
 public class SafariInfixSuggester extends AnalyzingInfixSuggester {
 
   private final boolean highlight;
-  private Set<BytesRef> suggestionSet;
+  private Map<Suggestion, Long> suggestWeightMap;
   private static final Logger LOG = LoggerFactory.getLogger(SafariInfixSuggester.class);
 
   public enum Context {
@@ -40,7 +39,7 @@ public SafariInfixSuggester(
 
     showContext = Collections.singleton(new BytesRef(new byte[] { (byte) Context.SHOW.ordinal() }));
     hideContext = Collections.singleton(new BytesRef(new byte[] { (byte) Context.HIDE.ordinal() }));
-    suggestionSet = new HashSet<>();
+    suggestWeightMap = new HashMap<>();
 
     if (!DirectoryReader.indexExists(dir)) {
       // no index in place -- build an empty one so we are prepared for updates
@@ -61,7 +60,7 @@ public void update(BytesRef bytes, long weight) throws IOException {
   public void build(InputIterator iter) throws IOException {
     // Reset suggestion HashSet on build
     LOG.info("\n\nStarting suggestion build.");
-    suggestionSet = new HashSet<>();
+    suggestWeightMap = new HashMap<>();
     super.build(iter);
   }
 
@@ -82,12 +81,32 @@ public void add(Dictionary dict) throws IOException {
     }
   }
 
-  // Override add method used during SuggestComponent suggest build to filter duplicates using HashSet.
+  /**
+   * Adds suggestion, only de-duplicating for the same text AND context,
+   * and keeping the duplicate with the highest weight.
+   * Calls AnalyzingInfix's add method if adding new suggestion,
+   * or AnalyzingInfix's update method if updating with a higher weight for existing suggestion.
+   * @param text BytesRef representing the text of suggestion
+   * @param contexts Set<BytesRef> representing the filter contexts for the suggestion
+   * @param weight the long weight of suggestion
+   * @param payload BytesRef payload of suggestion, usually used to store more metadata about suggestion
+   * @throws IOException
+   */
   @Override
   public void add(BytesRef text, Set<BytesRef> contexts, long weight, BytesRef payload) throws IOException {
-    if(suggestionSet.add(text)) {
+    Suggestion suggestion = new Suggestion(text, contexts, weight, payload);
+    Long currentSuggestWeight = suggestWeightMap.get(suggestion);
+
+    // Add suggestion if it has not yet been added.
+    if(currentSuggestWeight == null) {
+      suggestWeightMap.put(suggestion, weight);
       super.add(text, contexts, weight, payload);
     }
+    // If suggestion was already added with a lower weight, update suggestion with this weight
+    else if(currentSuggestWeight.doubleValue() < weight) {
+      suggestWeightMap.put(suggestion, weight);
+      super.update(text, contexts, weight, payload);
+    }
   }
 
   /*

src/main/java/com/ifactory/press/db/solr/spelling/suggest/Suggestion.java

@@ -0,0 +1,82 @@
+package com.ifactory.press.db.solr.spelling.suggest;
+
+import org.apache.lucene.util.BytesRef;
+
+import java.util.Set;
+
+public class Suggestion {
+
+  private String text;
+  private String payload;
+  private Set<BytesRef> contexts;
+  private long weight;
+
+  public Suggestion(BytesRef text, Set<BytesRef> contexts, long weight, BytesRef payload) {
+    this.text = text != null ? text.utf8ToString() : null;
+    this.contexts = contexts;
+    this.weight = weight;
+    this.payload = payload != null ? payload.utf8ToString() : null;
+  }
+
+  public String getText() {
+    return text;
+  }
+
+  public void setText(String text) {
+    this.text = text;
+  }
+
+  public String getPayload() {
+    return payload;
+  }
+
+  public void setPayload(String payload) {
+    this.payload = payload;
+  }
+
+  public Set<BytesRef> getContexts() {
+    return contexts;
+  }
+
+  public void setContexts(Set<BytesRef> contexts) {
+    this.contexts = contexts;
+  }
+
+  public long getWeight() {
+    return weight;
+  }
+
+  public void setWeight(long weight) {
+    this.weight = weight;
+  }
+
+  /**
+   * Overridden equals() method that only uses text and contexts variables for equality check.
+   * @param other the other Object to be checked for equality to this Suggestion
+   * @return true if 'other' is an instance of Suggestion and has the same text and
+   * contexts as this Suggestion object.
+   */
+  @Override
+  public boolean equals(Object other) {
+    if(this == other) {
+      return true;
+    }
+
+    if(!(other instanceof Suggestion)) {
+      return false;
+    }
+
+    Suggestion otherSuggestion = (Suggestion)other;
+
+    // Intentionally only using text and context to signify that two suggestions are equal.
+    return this.text.equals(otherSuggestion.text) && this.contexts.equals(otherSuggestion.contexts);
+  }
+
+  @Override
+  public int hashCode() {
+    int result = 17;
+    result = 31 * result + text.hashCode();
+    result = 31 * result + contexts.hashCode();
+    return result;
+  }
+}